Internet Drama

Internet Drama

Sometimes I can still be a superhero to my kids. For example, I can swiftly respond to this cry for help:

“Eeeek! There’s a spider in my room!”

I’ve handled this type of situation many times, and not to brag too much, but I’m pretty good at resolving this situation. I even have a special air gun that captures the bugs and allows me to set them free outside our house.

“Can you bring me a glass of water with ice?”

This type of request requires more agility than skill, as I need to descend and ascend steps and deliver the glass of water to one of my kids.

“Looks like the Internet is down.”

Like other cries for help, I’ve responded many times, and so I know how to fix this problem. I’ve been fixing Internet issues since my modem dial-up days over 30 years ago. With our new devices, it’s even easier. So when this cry for help came in two weeks ago, I simply unplugged the modem, waited ten seconds, and plugged it back in. The Internet, however, was still down.

Let’s call the cable company providing us Internet, Cable Company C. C has a fantastic process to troubleshoot and fix Internet problems. I went to their website and accessed their online chat service where I can ask any questions. After entering my phone number into the chatbot discussion, I learned that there are no outages in my area. That’s good to know. They also provided me with a phone number for their service department that I can text with and troubleshoot further.

I moved on to what I am assuming is Level 2 support. Although this text service is also automated (What happened to talking with a real person?), it was pretty sophisticated. After entering my phone number, it performed even more advanced troubleshooting and remotely rebooted my modem. Still no Internet. This texting automaton then forwarded our text conversation to a real person. The text conversation continued with this real person, who performed still more more advanced troubleshooting, and yet, no Internet. This Level 3 support person texted me a phone number I could call. I called Level 4 support, spoke with a skilled technician, who after accessing my account with my phone number, tried everything, and yet, still no Internet. He scheduled a technician to visit the next day to fix the problem. I can wait a day. Nobody came. I called up upset, the person apologized, assured me someone would be there the next day, and still, nobody came.

After four days of no Internet, and me spending hours on the phone and troubleshooting, the issue was finally resolved. Why did it take so long? After all, they had what appeared to be an amazing troubleshooting process.

The problem, I learned after, was a data issue. Or, to be more specific, a data modeling issue. (By the way, most issues in this world are data modeling issues. Just read the news headlines if you need proof.)

We live Arizona during most of the year and in New Jersey for the summer. I did not know that Cable Company C, which provides our Internet in New Jersey, and let’s call them Cable Company S, which provides our Internet in Arizona, are both owned by parent company A. The data model, therefore, should look something like this model, which I built with Ellie (learn more here: Ellie.ai):

No alt text provided for this image

Each Customer may own many Customer Locations. For example, Steve may own a location in Arizona and New Jersey. My phone number, which they asked me for each time I entered a new level of support, is how they identify me. That is, the Customer’s primary key. The primary key for Customer Location is both my phone number and some additional identifier, which could be an account number because they asked me for an account number that I did not know and therefore supply. There is more to the model above yet at a minimum, this model would allow the technicians to recognize that my phone number would bring up two locations and troubleshoot the correct location.

The problem was that they kept troubleshooting the wrong location! For example, whenever the modem would get rebooted, it was the Arizona modem getting rebooted instead of in New Jersey where I was having the outage. In fact, the technicians arrived at our Arizona location twice! The data model really looked more like this:

No alt text provided for this image

They have the concept of an Account that is specific to a Company. So, for example, Account Number 123 might exist in Cable Company C and also in Cable Company S, yet be two separate accounts. In addition, although my phone number was correctly tied to both of my accounts, only my first account came up and not the second (maybe appearing alphabetical?). I later learned that they even had my first name as “Steve” in Cable Company C and “Steven” in Cable Company S!

I have worked on many data integration projects over the years, and so I first thought the data issues were due to this being a new merger. Now that I had my Internet back, I did a quick search and learned that A bought C and S back in 2016! They should have their integration issues solved by now!

What’s even more frustrating is that their efficient processes fell apart because of poor data practices. As a result, they lost money and time, and frustrated a long-time customer. In fact, I have decided to leave both C and S at the end of the summer.

It is so important to get the data model right – everything builds on the data model. A sound design leads to working processes, great business decisions, and happy customers.

I’ll stick to getting rid of spiders and hope I never have another “bug” with my Internet!

Sanjeev Kumar Upadhyay

Founder Master Your Data Master Your Life?|Data Ki Baat?|Frame-works thinking to What-works thinking | Gang Of Data Stewards

2 年

What a great way of presenting a super important topic Steve Hoberman, DMC Loved your writing and storytelling.. anyone can read and understand no matter he/she data professional or not

回复
Larry Burns

Data and BI Architect at Fortune 500 Manufacturer

2 年

This brings up a point I made in "Data Model Storytelling": One of the most important things a data modeler does is to ask questions. Data modelers should always be asking "what if" questions to ensure that the data model is flexible and robust enough to meet every organizational need without breaking. The acquisition of the two cable companies by a third company should have triggered an immediate review of each company's data models, database structures and data. This obviously didn't happen.

Michel Hébert M.G.L. CDMP

Senior Consultant in Data Management

2 年

Great real-life example of the consequences of an incorrect data model... Thanks for sharing

Gwen Thomas

Independent voice helping organizations realize value from data and manage risk to it / from it. I’ve helped very large & mature groups and those just getting started. Founder of The Data Governance Institute (DGI)

2 年

Great story, Steve!

Marilu Lopez

Data Strategist and Data Management Consultant. Author of "Data Strategies for Data Governance"

2 年

Very good and illustrative story Steve Hoberman, DMC , thanks!

要查看或添加评论,请登录

Steve Hoberman, DMC的更多文章

  • Where are you from?

    Where are you from?

    My wife and I recently went on a cruise, and we chatted with many other passengers. We asked and answered many…

    16 条评论
  • No excuses

    No excuses

    I send a monthly email with data happenings that includes a personal story. (btw, if you don’t get these monthly…

    12 条评论
  • Hiking in the rain

    Hiking in the rain

    My wife, a neighbor, and I went on a hike a few weeks back. What started out overcast and a little drizzly turned into…

    1 条评论
  • Joe Versus the Volcano (or "Steve Versus the Data Scientist")

    Joe Versus the Volcano (or "Steve Versus the Data Scientist")

    My favorite movie is Joe Versus the Volcano, starring Tom Hanks and Meg Ryan. This 1990 movie didn’t get great reviews…

    14 条评论
  • "Two out of three ain't bad"

    "Two out of three ain't bad"

    Have you ever seen a triangle like this? Although this triangle has multiple names (triple constraint, iron triangle…

    17 条评论
  • Good data over good technology

    Good data over good technology

    Good data is more important than good technology. Take my 85-year-old neighbor, for example.

    16 条评论
  • Feeling the Data

    Feeling the Data

    My sailboat cuts through the choppy Barnegat Bay water, bumping up and down rhythmically over the small wave crests…

    22 条评论
  • The Power of Definitions

    The Power of Definitions

    The line looks way too long as I and many other weary red night warriors approach the end of the customs and…

    41 条评论
  • What would you climb a mountain for?

    What would you climb a mountain for?

    What would you climb a mountain for? We were each answering this question around our dinner table during a recent…

    7 条评论
  • To host an in-person conference, or not?

    To host an in-person conference, or not?

    I advanced to the final PowerPoint slide, and a multiple-choice question appeared on the large ballroom screen, “Where…

    31 条评论

社区洞察

其他会员也浏览了